Releases: MiaoY0uShan/FP
Releases · MiaoY0uShan/FP
Release list
FP v0.5.4
What's New
README rewritten with product narrative
- Problem → proof → action structure (Jobs launch + Belfort Straight Line method)
- Leads with the results users care about: 0 blockers, +11% routing accuracy, 4.61 goal fidelity
- Full ranking data in a collapsible block — nothing hidden, just prioritized
- Per-trait delta table (FP vs No FP) for transparency
- FAQ rewritten as objection handling
Principle vs hard-rule distinction in Four Iron Rules (四条铁律)
- Items 1–3 now labeled (原则) — default stance, not absolute prohibition
- Item 4 remains (硬规则) — no exceptions
- Philosophy: "默认立场，不是绝对禁令" — pragmatic deviation with stated reason is allowed
Team preset redesigned
- All discipline modules active by default with
## MODULE:headers - Each section has
> Remove this section if...guidance - No more commented-out blocks requiring manual uncommenting
Strategy tiers simplified
- Removed three-tier framing from README
- New guidance: "Default gives you everything. Remove what you don't need."
fp/presets/core.mdavailable for minimal setups
Reproducible benchmarks
benchmarks/run-eval.sh/run-eval.cmd— one-click full evalbenchmarks/README.md— complete reproduction documentation- Version snapshots for exact comparison
Benchmark Results (this session)
| Version | Score | Blockers |
|---|---|---|
| v-final (0.5.4) | 4.12 | 0 |
| v0 (no FP) | 4.10 | 3 |
| v-pre053 (absolute language) | 4.08 | 2 |
| v-core (rules only) | 4.05 | 2 |
Principles beat absolute commands (4.12 vs 4.08). Full discipline beats core-only (4.12 vs 4.05).
Full Changelog: v0.5.3...v0.5.4
FP v0.5.3
What's New
- English development discipline in README — added translations of 底层开发思维, 模块边界, 文档同步, 临时残留代码清理, and 四条铁律 to the English README
- Version bump across all 13 install packs and contracts
Full Changelog
Full Changelog: v0.5.2...v0.5.3
FP v0.5.2
What's New
- 开发纪律 (Development Discipline) — 4 个中文板块直接加入核心 SKILL.md：底层开发思维、模块边界、文档同步、临时残留代码清理
- fp cool 替代 Idea Cards — Vague 路由改用追问面试机制，用户说"问我问题"即可触发
- 移除 Reuse Ladder — 功能已被底层开发思维覆盖
Full Changelog: v0.5.1...v0.5.2
Full Changelog: v0.5.1...v0.5.2
FP v0.5.1
FP v0.5.0 — Evidence-Based Optimization
What changed
SKILL.md: 162 → 77 lines — templates 35 → 7, on-demand profiles, zero ceremony creep.
fp-minimal — 3-rule variant for non-reasoning models (gpt-5.3, etc.)
Evidence (1,416 real LLM API calls)
| Model | Winner | Score |
|---|---|---|
| gpt-5.3-codex-spark | v-minimal | 3.08 |
| gpt-5.6-sol | v-final (77 lines) | 3.57 |
| deepseek-v4-pro | v-final (77 lines) | 3.14 |
E2E multi-turn: 45% fewer tokens, 89% fewer FP template reads, correct on-demand profile triggering.
Templates: 28 dead templates archived. Agents no longer waste tokens reading them by default.
Key insights
- Less is more only on weak models. Reasoning models need structure.
- Simulation predicted v7 wins at 4.73. Real eval: v0 baseline won.
- Prompt interference is real: adding CRITICAL modifiers degrades performance.
Full report: benchmarks/results/ARTICLE.md
FP v0.4.10
Full Changelog: v0.4.9...v0.4.10
Full Changelog: v0.4.9...v0.4.10